The application of deep learning to X ray astronomy

Deep learning has been used to great effect in the classification of optical/lR galaxy images (e.g. arxiv.org/abs/1709.05834; adsabs.harvard.edu/abs/2017IAUS .. 325 .. 205C). By now this technique out performs Galaxy Zoo and has made human classifications
redundant. The student will apply the same approach to more than 10,000 X-ray images from the XMM-Newton satellite. Once refined on existing data, the student will develop analysis schemes that will maximise the science output of the forthcoming e-ROSITA and Athena missions. In addition to the data science aspects of this project, it will facilitate several astronomy studies. The resulting improved
knowledge of the cluster selection function will enhance both the precision of cosmological parameters extracted from cluster number counts, and understanding of the physics at play in the intracluster medium. The resulting improved fidelty of the non-cluster XMM source population will facilitate studies of rare objects, such as Isolated Neutron Stars, X-ray Flashes and distant quasars.